Elasticsearch: filter for a substring in the value of a document field?

Question

I am new to Elasticsearch. I have the following mapping for a string field:

"ipAddress": {
  "type": "string",
  "store": "no",
  "index": "not_analyzed",
  "omit_norms": "true",
  "include_in_all": false
}

A document with value in the ipAddress field looks like:

"ipAddress": "123.3.4.12 134.4.5.6"

Notice that in the above there are two IP addresses, separated by a blank.

Now I need to filter documents based on this field. This is an example filter value

123.3.4.12

And the filter value is always a single IP address as shown above.

I look at the filters at

http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/query-dsl-filters.html

and I cannot seem to be able to find right filter for this. I tried the term filter,

{
    "query": {
        "filtered" : {
            "query" : {
                "match_all" : {}
            },
            "filter": {
                "term" : { "ipAddress" : "123.3.4.12" }
            }
        }
    }
}

but it seems that it returns a document only when the filter value 100% matches the value of a document's field.

Can anyone help me out on this?

Update:

Based on John Petrone's suggestion, I got it working by defining a whitespace tokenizer based analyzer as follows:

{
  "settings": {
    "index": {
      "analysis": {
        "analyzer": {
          "blank_sep_analyzer": {
            "tokenizer": "whitespace"
          }
        }
      }
    }
  },
  "mappings": {
    "ipAddress": {
      "type": "string",
      "store": "no",
      "index": "analyzed",
      "analyzer": "blank_sep_analyzer",
      "omit_norms": "true",
      "include_in_all": false
    }
  }
}

Answer 1

The problem is that the field is not analyzed, so if you have 2 IP addresses in it the term is actually the full field, eg "123.3.4.12 134.4.5.6".

I'd suggest a different approach - if you are always going to have lists of IP addresses separated by spaces consider using the whitespace tokenizer to create tokens as whitespaces - should create several tokens that the IP address will then match:

http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/analysis-whitespace-tokenizer.html

Answer 2

Another approach could be storing the IP addresses as an array. And then the current mappings would work. You would just have to separate the IP addresses when indexing the document.